Paid app. Coordinate owned GPU leases, bounded CPU workers and optional resource reporting across compatible computers. Before you start: GPU admission requires supported NVIDIA hardware. Fleet work needs a trusted coordinator and explicit permissions; VRAM is not pooled.